Amazon product rankings are divided into many categories. Dresses has small category rankings. Among them, the Casual small category ranks 31. There is also a total product ranking in the large category, and Shops ranks 391 in the women’s clothing category. These rankings represent the The competitiveness and sales of listings on Amazon. The higher the ranking, the greater the competitiveness and the greater the sales.
The top 6 products in the Casual category in the Dresses category. So why is the product “DemetoryWomen's Summer Floral Print Sleeveless Scoop Neck Tunic Dress with Pocket” ranked first, while “AUSELILY Women&# Which product “39;s Sleeveless Pockets Casual Swing T-Shirt Dresses” ranks second? If we click on their links, we can see their respective rankings under the Shops category.
You can intuitively see that the reason why the second product is ranked second in the Casual category is determined by the ranking displayed on its page, and its small category ranking is determined by its large category ranking, that is, the shops ranking. The ranking of Shops is determined by its sales volume, from which the following logic can be obtained:
Sales → Large Category Ranking → Small Category Ranking.
So is the ranking directly linked to the keyword search ranking? Because there are many combinations of keywords, here we choose the simplest and most direct product root attribute for combination, that is, Casual + Dress, and search for CasualDresses in the Womer category.
Exclude advertisements and intercept the first 3 columns of products.
We can notice that the product that originally ranked second in the Casual sub-category only ranked 7th under this keyword search, and the product that originally ranked first in the Casual sub-category actually ranked 15th. , so we can get the second logic:
Category ranking ≠ keyword search ranking (even if the keyword completely matches the category).
So why doesn’t Amazon place the top-ranked products in the search field in a position that matches the category ranking? This is because there are thousands of keyword combinations, and each one The search groups corresponding to each combination are also completely different. A high-ranking product listing has a high probability of ranking high in multiple combinations, but it does not mean that each ranking can reach the top position. Each product corresponds to each keyword. The matching degree of the combination is different, which can be understood by referring to the manual advertising optimization matrix.
In addition to the above two logics, Amazon practitioners generally believe that only by optimizing the conversion rate can product rankings be improved. Their logic is (note: this logic is wrong!): Under the same traffic, a higher conversion rate means greater sales, and Amazon can also get more commissions, so the conversion rate is the key to keyword search rankings. decisive factor.
The above logic error is that it is only limited to the comparison of a single listing and ignores the customer purchase logic and platform operation logic.
The first is the customer purchasing logic. For a single user, shopping on Amazon requires a three-step cycle of search + browsing and one click, and finally purchases the product of their choice. Then we can deduce a logical conclusion: the probability of a customer’s transaction is directly proportional to his browsing time.
The second is the platform operation logic. Although the traffic is different for each listing, for the Amazon platform as a whole or a certain major category, the daily traffic tends to be a stable value. We can prove this point on Google Trends.
How can we maximize platform revenue? The purpose of Amazon’s operations is to try its best to enable every user who logs in to Amazon to complete orders, that is, Amazon will use various methods to make the conversion rate of flat combined orders closer At 1 (of course, the number of consecutive clicks by users is limited, so users must be prompted to purchase within the limited click recommendation period. Generally speaking, users will not click on recommended links more than 25 times in a row).
We have now reached two conclusions:
(1) The probability of a customer’s transaction is directly proportional to their browsing time.
(2) Amazon will use various methods to make the single traffic conversion rate of the platform close to 1. From this we can make the following mathematical deduction.
Suppose a customer browses n listings after searching for a certain keyword on Amazon. The time spent on the i-th listing is ti, and the total time spent. At the same time, the conversion rate of the i-th listing product is Pi. Assume The customer transaction probability is Y, then an equation can be established.
It can be seen that, when the coefficient, k, remains unchanged, the longer and higher the total time for users to browse the listing, the better, and at the same time, the conversion rate of the listing itself is also higher, the better.
Simulate the user’s browsing logic.
The meaning is that after the user browses the first listing, he will perform two actions: “jump out of the page” and “click on recommended products”, and then start browsing the second listing. Such actions will cycle. Until the user exits the Amazon website or completes the order.
As one of the links, our listing must participate in the ranking of the entire Amazon A9 algorithm. At this time, we need to analyze the different positions of the listing. It is divided into the following three situations:
(1) Head listing, that is, high keyword ranking listing
(2) Middle listing, that is, medium keyword ranking listing.
(3) Tail listing, that is, low keyword ranking listing.
There are two positioning ideas for the head listing: one is to serve as the traffic entrance, and the other is to serve as the traffic end point. Function. The meaning of traffic entrance is Amazon A9 algorithm evaluation. Although your listing conversion rate is not very competitive, because the exposure conversion rate of product pictures is higher and it shares more attributes with high-converting listings, it will be given more exposure. Higher keyword ranking.
Why do we need a higher exposure conversion rate? Because as a traffic portal, customers must be willing to click on this link. Otherwise, there will only be exposure and no traffic. Even if the conversion rate is high, the A9 algorithm It will not give this listing a high search ranking.
Why do you need to share more attributes with a high-converting listing? Because it is an entry point rather than an end point, the Amazon A9 algorithm does not expect traffic to complete orders on your listing. , but hope to be able to expose more high-converting listings through your listing recommendations to guide users to click and complete the purchase, thereby making the platform’s single traffic conversion rate approach 1. However, if there are no products with the same attributes as ours, Or other listings with similar attributes have very low conversion rates. So unless our listing can become a traffic destination and try our best to increase the conversion rate to complete the order, it is impossible to have a high search ranking.
Regarding the central listing, its positioning idea is to serve as a traffic transfer station, that is, to divert traffic from high-ranking listings to low-ranking listings.
Why do we need to divert traffic from high-ranking listings? Because it is impossible to achieve 100%. Converted product listings. So if there is a high-ranking listing with a low conversion rate, but it shares many attributes with your listing, and your listing has a relatively high conversion rate, then part of your listing’s traffic will be It will come from that high-ranking listing.
Why do you need to divert traffic to low-ranking listings? In the same way as the former, no matter how excellent your listing conversion rate performance is, it cannot achieve 100% conversion, so the A9 algorithm will continue to convert part of the traffic that has not been transacted to the conversion rate. Good lead for low-ranking listings.
As for the tail listing, its positioning idea is basically to let it assume the role of the end of the traffic cycle and restart the traffic cycle. The meaning of the end point of the traffic cycle is that regardless of whether the user will make a purchase at the end of the listing to generate an order, the Amazon A9 algorithm defaults to the end of the cycle and will no longer recommend products with the same attributes but generally only recommend new products under the same brand. At this time, because there are no shared attributes, users will basically exit the link and start searching for new keywords to enter the next traffic cycle. The meaning of restarting the traffic cycle is that in addition to directly exposing other new products in the store, the Amazon A9 algorithm may directly add a certain first listing as the most recommended link to allow users to enter the next traffic cycle.